A dataset of neonatal EEG recordings with seizure annotations.

Neonatal seizures are a common emergency in the neonatal intensive care unit (NICU). There are many questions yet to be answered regarding the temporal/spatial characteristics of seizures from different pathologies, response to medication, effects on neurodevelopment and optimal detection. The dataset presented in this descriptor contains EEG recordings from human neonates, the visual interpretation of the EEG by the human experts, supporting clinical data and codes to assist access. Multi-channel EEG was recorded from 79 term neonates admitted to the NICU at the Helsinki University Hospital. The median recording duration was 74 min (IQR: 64 to 96 min). The presence of seizures in the EEGs was annotated independently by three experts. An average of 460 seizures were annotated per expert in the dataset; 39 neonates had seizures and 22 were seizure free, by consensus. The dataset can be used as a reference set of neonatal seizures, in studies of inter-observer agreement and for the development of automated methods of seizure detection and other EEG analyses.

Functional maturation in preterm infants measured by serial recording of cortical activity.

Minimally invasive, automated cot-side tools for monitoring early neurological development can be used to guide individual treatment and benchmark novel interventional studies. We develop an automated estimate of the EEG maturational age (EMA) for application to serial recordings in preterm infants. The EMA estimate was based on a combination of 23 computational features estimated from both the full EEG recording and a period of low EEG activity (46 features in total). The combination function (support vector regression) was trained using 101 serial EEG recordings from 39 preterm infants with a gestational age less than 28 weeks and normal neurodevelopmental outcome at 12 months of age. EEG recordings were performed from 24 to 38 weeks post-menstrual age (PMA). The correlation between the EMA and the clinically determined PMA at the time of EEG recording was 0.936 (95%CI: 0.932-0.976; n = 39). All infants had an increase in EMA between the first and last EEG recording and 57/62 (92%) of repeated measures within an infant had an increasing EMA with PMA of EEG recording. The EMA is a surrogate measure of age that can accurately determine brain maturation in preterm infants.

A novel measure of reliability in Diffusion Tensor Imaging after data rejections due to subject motion.

Diffusion Tensor Imaging (DTI) is commonly challenged by subject motion during data acquisition, which often leads to corrupted image data. Currently used procedure in DTI analysis is to correct or completely reject such data before tensor estimations, however assessing the reliability and accuracy of the estimated tensor in such situations has evaded previous studies. This work aims to define the loss of data accuracy with increasing image rejections, and to define a robust method for assessing reliability of the result at voxel level. We carried out simulations of every possible sub-scheme (N=1,073,567,387) of Jones30 gradient scheme, followed by confirming the idea with MRI data from four newborn and three adult subjects. We assessed the relative error of the most commonly used tensor estimates for DTI and tractography studies, fractional anisotropy (FA) and the major orientation vector (V1), respectively. The error was estimated using two measures, the widely used electric potential (EP) criteria as well as the rotationally variant condition number (CN). Our results show that CN and EP are comparable in situations with very few rejections, but CN becomes clearly more sensitive to depicting errors when more gradient vectors and images were rejected. The error in FA and V1 was also found depend on the actual FA level in the given voxel; low actual FA levels were related to high relative errors in the FA and V1 estimates. Finally, the results were confirmed with clinical MRI data. This showed that the errors after rejections are, indeed, inhomogeneous across brain regions. The FA and V1 errors become progressively larger when moving from the thick white matter bundles towards more superficial subcortical structures. Our findings suggest that i) CN is a useful estimator of data reliability at voxel level, and ii) DTI preprocessing with data rejections leads to major challenges when assessing brain tissue with lower FA levels, such as all newborn brain, as well as the adult superficial, subcortical areas commonly traced in precise connectivity analyses between cortical regions.

Estimating functional brain maturity in very and extremely preterm neonates using automated analysis of the electroencephalogram.

OBJECTIVE: To develop an automated estimate of EEG maturational age (EMA) for preterm neonates. METHODS: The EMA estimator was based on the analysis of hourly epochs of EEG from 49 neonates with gestational age (GA) ranging from 23 to 32weeks. Neonates had appropriate EEG for GA based on visual interpretation of the EEG. The EMA estimator used a linear combination (support vector regression) of a subset of 41 features based on amplitude, temporal and spatial characteristics of EEG segments. Estimator performance was measured with the mean square error (MSE), standard deviation of the estimate (SD) and the percentage error (SE) between the known GA and estimated EMA. RESULTS: The EMA estimator provided an unbiased estimate of EMA with a MSE of 82days (SD=9.1days; SE=4.8%) which was significantly lower than a nominal reading (the mean GA in the dataset; MSE of 267days, SD of 16.3days, SE=8.4%: p<0.001). The EMA estimator with the lowest MSE used amplitude, spatial and temporal EEG characteristics. CONCLUSIONS: The proposed automated EMA estimator provides an accurate estimate of EMA in early preterm neonates. SIGNIFICANCE: Automated analysis of the EEG provides a widely accessible, noninvasive and continuous assessment of functional brain maturity.

Early development of synchrony in cortical activations in the human.

Early intermittent cortical activity is thought to play a crucial role in the growth of neuronal network development, and large scale brain networks are known to provide the basis for higher brain functions. Yet, the early development of the large scale synchrony in cortical activations is unknown. Here, we tested the hypothesis that the early intermittent cortical activations seen in the human scalp EEG show a clear developmental course during the last trimester of pregnancy, the period of intensive growth of cortico-cortical connections. We recorded scalp EEG from altogether 22 premature infants at post-menstrual age between 30 and 44 weeks, and the early cortical synchrony was quantified using recently introduced activation synchrony index (ASI). The developmental correlations of ASI were computed for individual EEG signals as well as anatomically and mathematically defined spatial subgroups. We report two main findings. First, we observed a robust and statistically significant increase in ASI in all cortical areas. Second, there were significant spatial gradients in the synchrony in fronto-occipital and left-to-right directions. These findings provide evidence that early cortical activity is increasingly synchronized across the neocortex. The ASI-based metrics introduced in our work allow direct translational comparison to in vivo animal models, as well as hold promise for implementation as a functional developmental biomarker in future research on human neonates.

The effect of fontanel on scalp EEG potentials in the neonate.

OBJECTIVE: To define how fontanels affect scalp EEG potentials in neonates. METHOD: Realistic finite element method head models were generated with and without fontanels. The electrical activity of the whole cortex was simulated using distributed 54,620 concurrently active cortical dipoles with a uniform random distribution of current densities (0-40 µA/cm2). The overall effects of fontanels on scalp potentials were calculated from finite element forward solution in the vicinity the fontanel region by relative difference measure (RDM*) and magnification factor (MAG), and the skull conductivity was systematically varied from 0.003 to 0.3S/m. RESULTS: The neonatal scalp EEG topographies are comparable in models with and without fontanels, with highest amplitudes directly above the anterior fontanel. Quantitatively, comparison of these models elicits negligible differences (RDM*) ∼2% and MAG ∼1.0). However, fontanel contribution to scalp potential was shown to increase sharply with decreases in skull conductivity. CONCLUSIONS: These results suggest that fontanels may affect neonatal scalp EEG much less than traditionally assumed, and the effect is strongly dependent on skull conductivity. SIGNIFICANCE: Most neonatal EEG studies can be adequately performed without assuming distortions by fontanels.

Measuring spike strength in patients with continuous spikes and waves during sleep: comparison of methods for prospective use as a clinical index.

OBJECTIVE: To compare methods of estimating spike strength as a potential index in the assessment of continuous spikes and waves during sleep (CSWS). METHODS: Spikes were searched and averaged automatically from pre- and postoperative EEGs of ten patients with CSWS who underwent corpus callosotomy (eight) or resective epilepsy surgery (two). From the most prominent spike, we measured peak amplitude and root mean square (RMS) over ±150ms window around the peak. In order to compensate for spatiotemporal instability of spikes, the cumulative amplitude and RMS were computed from the highest quartile of electrodes (Ampl-Q and RMS-Q, respectively). The stability of parameters was studied by comparing two ten minute epochs during the first hour of NREM sleep, as well as by analyzing overnight variation of indices in further ten patients with CSWS. The Ampl-Q and RMS-Q were compared between pre- and postoperative recordings. RESULTS: All four measures, amplitude, RMS, Ampl-Q and RMS-Q, were correlated with each other and highly dependent on NREM/REM-sleep stage and arousals. Expectedly, Ampl-Q and RMS-Q had the greatest intra-individual stability. The amplitude had up to 71% intra-individual variation making it unhelpful for clinical use. Ampl-Q and RMS-Q were comparable in assessing change following surgical treatment. CONCLUSIONS: Computing an integrated RMS over multiple electrodes during steady NREM sleep offers a stable and reliable parameter for evaluating the strength of spikes in CSWS. SIGNIFICANCE: Analyzing spike strength with RMS-Q may offer a clinically useful, supplementary index for EEG monitoring of CSWS where spike index has been of limited value.

Early neurophysiology and MRI in predicting neurological outcome at 9-10 years after birth asphyxia.

OBJECTIVE: To assess whether early somatosensory evoked potentials (SEP) predict long-term neurodevelopmental outcome in normothermic, full-term infants with mild to moderate neonatal encephalopathy (NE), and to compare their predictive value to already available amplitude integrated EEG (aEEG) and magnetic resonance imaging (MRI). METHODS: Fifty-six infants with post-asphyxia NE were prospectively recruited, and their SEP, aEEG and MRI data were acquired during the first five days. Follow-up continued to 9-10 years for assessment of neuromotor and neurocognitive development. We analysed SEP latency (N1 component), normality of aEEG background pattern, as well as patterns of injury on the neonatal MRI. Neurological outcome measures at 9-10 years included conventional MRI, Movement-ABC and the WISC-III NL. RESULTS: A SEP latency <50 ms during the first five days was associated with a normal neuromotor outcome (p < 0.03), and a prolonged day 3 latency was associated with lower childhood IQ (p = 0.02). The presence of multiple seizures in aEEG, as well as a moderate or severe injury on the neonatal MRI was associated with a poor neuromotor score (p = 0.03 and p < 0.01, respectively). Combination of multiple techniques improved prediction of long-term outcome compared to single modality. CONCLUSION: Early SEPs provide information that is comparable to the already available aEEG and MRI paradigms in the prediction of long-term outcome of full-term infants with mild to moderate neonatal encephalopathy. SIGNIFICANCE: The present results call for further studies using early SEP to aid early assessment of infants treated with hypothermia.

Semiautomatic quantification of spiking in patients with continuous spikes and waves in sleep: sensitivity to settings and correspondence to visual assessment.

OBJECTIVE: To define the optimal analysis protocol for semiautomatic quantification of spike index (SI) in continuous spikes and waves in sleep (CSWS). METHODS: Ten overnight EEGs (nine patients) with abundant spiking were used to quantify SI with a previously published semiautomatic quantification based on spike detection with BESA software. We studied (i) dependency of SI on maximal interspike interval (maxISI) defining the continuous discharge, (ii) sensitivity of SI to variations in the spike search protocol, and (iii) stability of SI over time. Finally, the semiautomatic method was compared with the quantification based on visual scoring by two neurophysiologists. RESULTS: MaxISI of 3s appeared to yield the best combination of sensitivity and stability in SI quantification. The SI of the first hour of sleep did not differ significantly from the SI of the whole night. Mean error of the semiautomatic method compared to visual scoring was only seven percentage units. CONCLUSIONS: Semiautomatic quantification of SI functions well with maxISI of 3s, and the first hour of sleep represents the whole night SI with a clinically relevant accuracy. SIGNIFICANCE: This method opens a possibility for objective quantification of near-continuous epileptiform spiking during sleep, and it supports the use of shorter epochs for quantitative assessment of CSWS.

Efficient reduction of stimulus artefact in TMS-EEG by epithelial short-circuiting by mini-punctures.

OBJECTIVE: We aimed at comparing the effects of two different electrode-to-skin contact preparation techniques on the stimulus artefact induced by transcranial magnetic stimulation (TMS) in electroencephalography (EEG) signals. METHODS: Six healthy subjects participated in a combined navigated brain stimulation (NBS) and EEG study. Electrode contacts were first prepared in the standard way of rubbing the skin using a wooden stick with a cotton tip. The location of hand motor area and the motor threshold (MT) was determined for each subject. Then, the TMS-induced artefact was measured at 60%, 80%, 100% and 120% of the MT. Subsequently, the epithelium under the electrode contacts was electrically short-circuited by puncturing with custom-made needles and the stimulation sequences were replicated. The artefact was compared between the preparation techniques. RESULTS: The TMS-induced artefact was significantly reduced after puncturing. In addition, the size and duration of the artefact depended on the applied stimulation intensity. The reduction of the artefact was largest in electrodes at and close to the stimulation site. CONCLUSIONS: Mini-puncturing technique enables more accurate analysis of TMS-induced short-latency phenomena in EEG during NBS, and it may aid in the examination of the short distance neural connectivity beneath and close to the stimulation site. SIGNIFICANCE: This study describes a practical skin preparation method that significantly improves the utility of TMS-EEG method in studying short-latency cortical connectivity.

Development of the spontaneous activity transients and ongoing cortical activity in human preterm babies.

Recent experimental studies have shown that developing cortex in several animals species, including humans, exhibits spontaneous intermittent activity that is believed to be crucial for the proper wiring of early brain networks. The present study examined the developmental changes in these spontaneous activity transients (SAT) and in other ongoing cortical activities in human preterm babies. Full-band electroencephalography (FbEEG) recordings were obtained from 16 babies at conceptional ages between 32.8 and 40 wk. We examined the SATs and the intervening ongoing cortical activities (inter-SAT; iSAT) with average waveforms, their variance and power, as well as with wavelet-based time-frequency analyses. Our results show, that the low frequency power and the variance of the average waveform of SAT decrease during development. There was a simultaneous increase in the activity at higher frequencies, with most pronounced increase at theta-alpha range (4-9 Hz). In addition to the overall increase, the activity at higher frequencies showed an increased grouping into bursts that are nested in the low frequency (0.5-1 Hz) waves. Analysis of the iSAT epochs showed a developmental increase in power at lower frequencies in quiet sleep. There was an increase in a wide range of higher frequencies (4-16 Hz), whereas the ratio of beta (16-30 Hz) and theta-alpha (4-9 Hz) range activity declined, indicating a preferential increase at theta-alpha range activity. Notably, SAT and iSAT activities remained distinct throughout the development in all measures used in our study. The present results are consistent with the idea that SAT and the other ongoing cortical activities are distinct functional entities. Recognition of these two basic mechanisms in the cortical activity in preterm human babies opens new rational approaches for an evaluation and monitoring of early human brain function.

Evaluation of commercially available electrodes and gels for recording of slow EEG potentials.

OBJECTIVE: To test the applicability of different types of commercially available electrodes and electrode gels or pastes for recording of slow EEG potentials. METHODS: Experiments were carried out on six types of reusable electrodes (silver, tin and gold cup electrodes, sintered silver-silver chloride (Ag|AgCl), platinum, stainless steel), six disposable Ag|AgCl electrode models, and nine gels or pastes. We studied the parameters, which are critical in slow-potential recording, such as polarization, initial and long-term stability and low-frequency noise. RESULTS: The best results were obtained with the reusable sintered Ag|AgCl electrodes. The six disposable Ag|AgCl electrode models also proved to have appropriate electrical properties. Other types of reusable electrodes suffered from diverse degrees of polarization, baseline drift, low-frequency noise, high resistance, and changes in properties due to wear and tear. Seven out of nine gels or pastes contained a significant amount of chloride, which is a prerequisite for DC stability of Ag|AgCl electrodes, whereas the absolute concentration of chloride had little effect. CONCLUSIONS: Direct current (DC) coupled recording of EEG is critically dependent on the choice of electrode and gel. SIGNIFICANCE: Our results provide rigorous criteria for choosing DC-stable electrodes and gels for DC-coupled or long time-constant AC-coupled recordings of slow EEG potentials.

Cerebrospinal fluid insulin-like growth factor-1, insulin growth factor binding protein-2 or nitric oxide are not increased in MS or ALS.

OBJECTIVES: Many studies have shown that nitric oxide (NO) and growth factors including insulin growth factors (IGFs) may be involved in the pathogenesis of multiple sclerosis (MS) and neurodegenerative diseases. Our previous studies suggested a relationship between cerebrospinal fluid (CSF) NO metabolites (nitrates and nitrites, NN(x)) and IGF-1 in patients with progressive encephalopathy, hypsarrhythmia and optic atrophy syndrome. MATERIAL AND METHODS: We examined CSF concentrations of NN(x), IGF-1 and IGF binding protein-2 (IGFBP-2) in 25 controls, 14 patients with MS and 14 patients with amyotrophic lateralis sclerosis (ALS). RESULTS: There were no significant differences in CSF levels of NN(x), IGF-1 or IGFBP-2 between the groups. CSF IGFBP-2 concentrations correlated significantly with age in controls, which may reflect age-related changes in the blood-brain barrier function. CONCLUSION: Upregulation of the production of NO and IGF-1 in the brain or spinal cord does not influence CSF levels of these molecules in MS or ALS.

Infraslow oscillations modulate excitability and interictal epileptic activity in the human cortex during sleep.

Human cortical activity has been intensively examined at frequencies ranging from 0.5 Hz to several hundred Hz. Recent studies have, however, reported also infraslow fluctuations in neuronal population activity, magnitude of electroencephalographic oscillations, discrete sleep events, as well as in the occurrence of interictal events. Here we use direct current electroencephalography to demonstrate large-scale infraslow oscillations in the human cortex at frequencies ranging from 0.02 to 0.2 Hz. These oscillations, which are not detectable in conventional electroencephalography because of its limited recording bandwidth (typical lower limit 0.5 Hz), were observed in widespread cortical regions. Notably, the infraslow oscillations were strongly synchronized with faster activities, as well as with the interictal epileptic events and K complexes. Our findings suggest that the infraslow oscillations represent a slow, cyclic modulation of cortical gross excitability, providing also a putative mechanism for the as yet enigmatic aggravation of epileptic activity during sleep.

Oral treatment of acute myeloid leukaemia with etoposide, thioguanine, and idarubicin (ETI) in elderly patients: a prospective randomised comparison with intravenous cytarabine, idarubicin, and thioguanine in the second and third treatment cycle.

A randomised multicentre study was conducted among patients over 65 yr of age with newly diagnosed acute myeloid leukaemia (AML) to compare oral treatment with etoposide 80 mg/m(2) and thioguanine 100 mg/m(2) twice daily on 5 d and idarubicin 15 mg/m(2) on 3 d (ETI) to a mainly i.v. combination of cytarabine 100 mg/m(2) twice daily on 5 d, idarubicin 12 mg/m(2) x 1, and thioguanine (TAI). Ninety-two patients were enrolled. Their median age was 72 yr, range 65-84 yr. Sixty-five patients had de novo AML, 21 AML subsequent to myelodysplastic syndrome, and six treatment-related AML. They received at first a 6-d i.v. treatment with cytarabine and idarubicin. After the first treatment, 68 patients were randomised to receive two cycles of ETI (n = 36) or TAI (n = 32) and thereafter maintenance with mercaptopurine and methotrexate. Of the 92 patients, 52 (57%) achieved remission at some stage. The median survival was 10 months. There were no significant differences between the patients randomised to ETI or TAI in the remission rate (67% vs. 72%), survival (12 months from randomisation in both arms), event-free survival or relapse rate. The patients randomised to receive ETI spent significantly fewer days at hospital during the two randomised cycles (20 vs. 41 d, P = 0.010), and they had fewer days with infusions, shorter neutropenias and thrombocytopenias and fewer and less severe infections. In conclusion, treatment with oral ETI resulted in a similar antileukaemic effect as obtained with mainly i.v. TAI, with less toxicity and reduced need for hospitalisation.

Topography and elimination of slow EEG responses related to tongue movements.

This study aimed to characterize the topography and mechanism of slow EEG potentials related to tongue movement (TMP), and to devise a method for their elimination. Eight adult subjects were recorded with 8-channel direct current (DC) EEG (n = 5) or with a 256-channel, dense array EEG (n = 4). Subjects were requested to push their tongue toward the incisors in a closed mouth. We examined the time course and topography of the ensuing TMP. The underlying mechanism and elimination of TMP were studied by electrical isolation of the tongue with a latex sheet, and by short-circuiting of tongue to gingival pouch. Forward movement of the tongue caused global changes in scalp potentials, with frontal areas more positive, and the strongest gradients between the mastoid region and other scalp areas. Scalp current source density was highest near ear canals and orbital fossae. Electrical isolation of the tongue tip with a latex sheet resulted in a near-complete elimination of TMPs, while short-circuiting resulted in high amplitude responses, suggesting that TMPs are caused by a change in the geometry of conductive pathways. Our results indicate that significant scalp potentials are caused by even modest tongue movements, which may occur subconsciously during various cognitive tasks. The topography and small amplitude of TMPs make their off-line rejection difficult, while an isolation method of the kind used here may substantially diminish TMP artifacts in studies on slow cognitive potentials.

Scalp-recorded slow EEG responses generated in response to hemodynamic changes in the human brain.

OBJECTIVE: To study whether hemodynamic changes in human brain generate scalp-EEG responses. METHODS: Direct current EEG (DC-EEG) was recorded from 12 subjects during 5 non-invasive manipulations that affect intracranial hemodynamics by different mechanisms: bilateral jugular vein compression (JVC), head-up tilt (HUT), head-down tilt (HDT), Valsalva maneuver (VM), and Mueller maneuver (MM). DC shifts were compared to changes in cerebral blood volume (CBV) measured by near-infrared spectroscopy (NIRS). RESULTS: DC shifts were observed during all manipulations with highest amplitudes (up to 250 microV) at the midline electrodes, and the most pronounced changes (up to 15 microV/cm) in the DC voltage gradient around vertex. In spite of inter-individual variation in both amplitude and polarity, the DC shifts were consistent and reproducible for each subject and they showed a clear temporal correlation with changes in CBV. CONCLUSIONS: Our results indicate that hemodynamic changes in human brain are associated with marked DC shifts that cannot be accounted for by intracortical neuronal or glial currents. Instead, the data are consistent with a non-neuronal generator mechanism that is associated with the blood-brain barrier. SIGNIFICANCE: These findings have direct implications for mechanistic interpretation of slow EEG responses in various experimental paradigms.

Very slow EEG responses lateralize temporal lobe seizures: an evaluation of non-invasive DC-EEG.

BACKGROUND: This study tested the idea that very slow EEG responses (direct current [DC] potential shifts) could be detected noninvasively during temporal lobe (TL) seizures, and that these shifts give lateralizing information consistent with that obtained by other methods. METHODS: Seven patients with TL epilepsy (TLE) were recorded with scalp DC-EEG technique at bedside. All recordings were performed simultaneously with conventional EEG (scalp in five, and intracranially in two; two patients with scalp recordings were recorded intracranially later). Seizures in five patients originated in the mesial TL. Ictal DC shifts were evaluated by comparing them to the temporal evolution of ictal discharges, and by comparing the laterality of these shifts to the side of seizure onset defined by routine EEG and other presurgical diagnostic tests. RESULTS: All seizures (35/35) were associated with negative DC shifts at temporal derivations (30 to 150 micro V relative to vertex), beginning at the electrical seizure onset, and lasting for the whole seizure. In eight seizures (five patients) with documented mesial TL onset, the polarity of the DC shift was initially positive followed by a negative one after lateral spread of seizure activity. In all cases, the side of the EEG shift agreed with other diagnostic tests, and, at times, was more clearly lateralized than the conventional scalp EEG. CONCLUSIONS: DC-EEG recordings are practical and achievable at the bedside. Ictal DC shifts are consistently observed in scalp recordings in TL seizures, and reliably lateralize them. This method may hold promise in reducing the need for invasive monitoring in patients with TLE where other noninvasive tests are equivocal.

Dutch patients with progressive encephalopathy with edema, hypsarrhythmia, and optic atrophy (PEHO) syndrome.

PEHO syndrome (progressive encephalopathy, peripheral edema, hypsarrhythmia, and optic atrophy) is a neurodegenerative disorder first characterized in Finnish patients. Subsequent reports have occasionally identified cases of PEHO occurring in some other countries. We describe two Dutch children who represent the first reported cases with PEHO syndrome from western Europe. Both of these children showed typical characteristics of the syndrome, including generalized hypotonia and edema of their faces and extremities, profound psychomotor retardation, progressive cerebellar atrophy, and severe epilepsy which initially started as infantile spasms. Our experience has shown that distinguishing cases with true PEHO from those with other, clinically similar disorders requires a firm demonstration of both the presence and the progress of typical neuroradiological findings. Our present cases suggest that 1) PEHO syndrome is not limited to Finnish heritage, and that 2) many more patients with PEHO could be identified with increased knowledge of the syndrome.

Reversed visual field constrictions in children after vigabatrin withdrawal--true retinal recovery or improved test performance only?

While vigabatrin-associated visual field constrictions have been generally considered irreversible, some case reports have raised the hope of partial improvement after drug withdrawal in occasional patients. Here we describe seven children with epilepsy, whose visual field constrictions, as demonstrated by the kinetic perimetry (Goldmann), attenuated or recovered after discontinuation of vigabatrin therapy. While this improvement may be largely due to better performance in later test sessions, we want to raise the possibility that some visual field recovery may be possible at least in young patients.